Key: Binary Outcome, Single Group

(1) ELECT. The proportion of voters favoring candidate A, p = 55 / 100 = .55 = 55%. A 95% CI for P = 44.8%, 64.9% by the Fleiss quadratic method calculated by EpiTable. Since 50% falls within this interval, we cannot say with 95% confidence that victory is imminent.

(2) BREASTCA. If the null hypothesis were true, the number of cases in the sample would follow a binomial distribution with n = 1000 and P = .02. The expected number of cases (�) = np = (1000)(.02) = 20. The observed proportion, p = 32 / 1000 = .032. We test H0: p = .02 vs. H1: p not = .02. Using the EpiTable's binomial procedure, the two-sided p-value = .012. This provides evidence to reject H0.

(3) PREGRATS. p-value = Pr(X >= 12| H0 true) = .0011. Conclude: the malformation rate is significantly greater than expected.

(4) SMOKE.REC. Observed recidivism risk, p = 201 / 234 = .859. A 95% confidence interval for P = .808, .901 (via Epi Info's FREQ command. Data make it clear that te large percentage of subjects resume smoking within a year.

(5) BINSIZE. For d = .1, n = 96. For d = .05, n = 384.

(6) EDENTITION. p = 20 / 262 = .076; 95% CI for P = (.048, .117) (Fleiss quadratic method)